Find Assets In GenAsset

At some point your GenAsset workspace stops being "a few experiments" and becomes a library. That's when you need Find Assets In GenAsset, the pack's search node. Give it a keyword and it queries your workspace's assets, returning the matches as JSON plus a ready-to-use CSV of asset IDs.

It's the node for the "which asset was that?" moment - the one where you vaguely remember a project name, a tag, or a phrase from a prompt. The search runs against name, tags, and prompt text, so the memory that unlocks it doesn't have to be precise. And because it's a node, not a web page, you can wire the results into the rest of your workflow: search for assets matching a tag, pull the ids, and feed them into a load node automatically.

How it works

It's a GET to api/v1/assets with page, page_size, and an optional search parameter built from your search_query. The server returns a paginated list of assets; the node compacts each row down to the useful fields (id, name, version count, and the rest of the asset card) and hands you two things at once:

• assets_json - the compact asset list as formatted JSON, ready to inspect or parse.
• asset_ids_csv - the same ids joined into a comma-separated string, so you can copy them straight into a load node's asset_id field or split them in your own logic.

The search is paginated, with page (1–9999) and page_size (1–64, default 20). The tooltip is worth remembering: "Empty lists recent assets." If you leave search_query empty, the node becomes a "what's in my workspace" listing rather than a search - a handy way to browse.

Inputs and outputs that matter

Five required inputs, but only two you'll touch: search_query (leave empty for recent assets) and page_size (bump it if you want more than 20 at a time). base_url and token stay at their defaults, and page only matters past the first screen of results.

Outputs are assets_json, asset_ids_csv, and status_json (which includes the result count and echoes your query). For a beginner, the practical move is: run it with a tag, read assets_json, copy an id from asset_ids_csv into a load node.

Installation

Part of the steliosot/ComfyUI-GenAsset pack. ComfyUI Manager → search GenAsset → install, or:

cd ComfyUI/custom_nodes
git clone https://github.com/steliosot/ComfyUI-GenAsset.git

Restart after. No extra dependencies; needs a workspace token (widget, GENASSET_WORKSPACE_TOKEN, or ComfyUI/user/genasset.json).

Common issues

• Search returns nothing for a term you know is in a prompt - the search may index name/tags and current prompt fields but not necessarily every historical version's prompt. Try a broader term, or just browse recent assets.
• asset_ids_csv is empty but the workspace isn't - pagination. If you're past page 1 of many results, the ids you're looking for may simply be on another page.
• It's an output node, so it doesn't chain - Find Assets In GenAsset is marked as an output node, meaning it's a terminal in the graph's sense. You read its outputs and wire the strings onward, but you won't hang more generation nodes off it.

This is a "find your stuff" node, not a "generate stuff" node, and that's a compliment. Every library needs a search bar, and this one plugs straight into the rest of the workflow.